The Permit

Coase's insight was that an externality is a missing market — a cost uncounted because no one owns the thing at issue. The boldest cure follows directly: if the trouble is a missing market, then make one. Create property in the right to pollute, cap the total at what the world can bear, and let that right be bought and sold — and a price for the harm will emerge, not from a government's estimate, but from the trading of those who know their own costs best. This is the permit, and it is Coase's theorem turned into policy.

The design, proposed by John Dales and others and now widespread, is called cap-and-trade, and its two parts are in the name. First the cap: the authority decides the total quantity of a pollutant the environment can tolerate — so many tons of sulfur, of carbon — and issues exactly that many permits, each conferring the right to emit one unit. The total harm is thereby fixed, capped at a tolerable level by fiat. Then the trade: the permits may be bought and sold freely, so that a firm which can cut its emissions cheaply will do so and sell its spare permits, while a firm for which cutting is dear will buy permits instead. The permits flow to where they are most valued, the cuts happen where they are cheapest, and a market price for a ton of pollution emerges from the trading — the price of the externality, discovered rather than decreed.¹

The beauty of the scheme is that it achieves a given reduction at the least possible cost, and does so without anyone needing to know what that least-cost path is. A regulator ordering each firm to cut by the same amount would force expensive cuts on firms that could only abate dearly, while leaving cheap cuts unmade at firms that could have abated for a song. The market in permits avoids this: by letting firms trade, it concentrates the cutting wherever cutting is cheapest, equalizing the cost of the last ton abated across all firms, which is exactly the condition for cutting a given total at minimum cost. The authority need not know any firm's costs; it sets only the cap, and the trading sorts out the rest. It is the Coasean promise made good — a market in the right, finding the efficient allocation through trade — but with the cap supplied collectively, since the multitude who breathe the air could never have bargained for it themselves.²

The permit and the Pigovian tax are close cousins — both put a price on the harm — but they differ in a way that matters under uncertainty, a difference Martin Weitzman made precise. The tax fixes the price of the harm and lets the quantity adjust; the permit fixes the quantity and lets the price adjust. If we knew everything, the two would be equivalent — set the tax at the price the cap would have produced, and you reach the same place. But we do not know everything, and which instrument is safer depends on what we most fear getting wrong. If the damage from a little extra pollution rises sharply past some threshold — a climate tipping point, a toxic concentration — then controlling the quantity matters most, and the permit, which fixes it, is the safer choice. If instead the cost of cutting is what might spike — if an unexpectedly dear round of abatement could cripple the economy — then controlling the price matters most, and the tax, which fixes it, is safer. Prices or quantities: the choice between the tax and the permit is, at bottom, a choice about which uncertainty to bear.³

The permit market inherits the difficulties of its parts and adds some of its own. The cap must still be set, and setting it demands the same knowledge of the harm that pricing the tax did — too loose a cap and the permits are worthless and the harm unchecked, too tight and the cost is needless; the hard question of how much harm to allow does not vanish, it only moves from the tax rate to the cap. The permits, once created, are valuable, and how they are first handed out — auctioned to the highest bidder, or given free to existing polluters — is a distribution of wealth that can dwarf the efficiency at stake, and that the incumbents will fight to capture. And the whole edifice rests on measurement: a ton emitted must be a ton counted, or the market trades in fictions; the scheme demands a machinery of monitoring as real as the pollution it prices. The market in the harm is only as honest as the meter that measures the harm.⁴

For all that, the permit market is the clearest vindication of Coase's vision, and a genuine achievement of the field. It took the abstract claim — that an externality is a missing market and the cure is to define a tradeable right — and built the market in fact, complete with property, price, and exchange, in a thing that had never been owned: the right to put a ton of something into the common air. Where it has been well designed, as in the program that cut American acid rain at a fraction of the predicted cost, it has worked as the theory promised, finding the cheap reductions the regulators could never have found themselves. It is the externality internalized not by the state pricing the harm, as Pigou would, but by the state creating the property and the market pricing the harm — Coase's bargain, scaled up by collective decree to a multitude that could not bargain for itself.⁵

The permit is the missing market made on purpose: property invented in the right to pollute, a cap to fix the total harm, and a price for that harm found by trading it among those who bear the cost of abating it. It is the most literal answer to Coase — not to tax the externality, but to own it — and where it works, it internalizes the harm as cleanly as theory could wish. With the tax, the bargain, and the permit before us, the volume has surveyed the great cures for the externality, each a way of bringing the outside cost inside the price. It remains only to ask what the externality is, that all these cures should take the form of pricing, taxing, or owning it — and the answer, the flagship will show, lies in the nature of the price's ledger itself, and in the boundary, drawn not by nature but by property, between what it counts and what it leaves out.